This paper describes the construction and stylometric analysis of a parallel corpus of Japanese and Okinawan translations of the Gospel of John, focusing on the Edo and Meiji periods. Although there have been many studies on biblical translation in Japan, there has yet to be a unified parallel corpus. This study presents an overview and design of an online corpus tool that includes the digital texts of the Gospel of John in the Gützlaff’s, Bettelheim’s (Okinawan), Hepburn and Brown’s, Meiji Motoyaku, Steichen and Takahashi’s translations of the 19th century and Nicolai’s, Raguet’s, Taishō Revised after the Meiji period, Nagai’s, and Sakon’s translations of the early 20th century. This paper presents an overview and design of the online corpus tool. Here, Omeka S is used to output the corpus data according to Linked Open Data (LOD). These corpus data are then subjected to a quantitative analysis of stylistic similarities and influences between translations using stylometric analysis in the R language package called stylo. This study is a step toward a deeper understanding of the influence of biblical translation on the formation of the modern Japanese language. It is intended as a contribution to the digital humanities.

View full abstract

[in Japanese]

View full abstract

This study investigated the use of the adverb of degree “totemo” and its synonyms, “kekkou” and “sugoku”, by learners and native speakers of Japanese using “International Corpus of Japanese as a Second Language”. The results of a classification tree analysis using three variables as explanatory variables: level (beginner, intermediate, advanced, native speaker), situation (four types of tasks), and learning environment (domestic, foreign, native speaker) showed that learner level had the strongest effect on the tendency to use different adverbs. The results also revealed that intermediate and advanced learners and native speakers have different tendencies to use adverbs depending on the situation, while beginner learners have different tendencies to use adverbs depending on the learning environment. Furthermore, although the results for the learners' results approached the results for the native speakers at higher levels, the frequent use of “sugoku” identified at the advanced level suggested the need to provide information on refraining from the use of certain words in certain situations.

View full abstract

This paper introduces the GLAD database, an Anglicism database of which the author is involved in the development, and presents its current state and prospective use. The GLAD database is open data developed by the cooperation of scholars studying Anglicisms in 17 different languages, with the aim to help promote cross-linguistic comparisons. This paper describes the background, criteria, and methods used for the development of the database, as well as issues such as unavoidable inconsistences among the datasets of the 17 languages. It also introduces the potential usage of the database in order to promote the use of the database and help progress the research of global Anglicisms.

View full abstract

A statistical analysis of the social acceptability of COVID-19 infection-related terms was conducted. The data were collected by the Agency for Cultural Affairs in 2021 as part of the “Opinion Survey on the Japanese Language”. The method of analysis was logistic regression analysis, with the explanatory variables being year of birth (or age) and gender, and the objective variable being social acceptability. The results of the analysis showed that the effects of year of birth (or age) and gender were significant for katakana words, and only the effect of gender was significant for abbreviated words. On the other hand, neither the effects of year of birth (or age) nor gender were significant for the four-character kanji phrases. Some discussion of the results was attempted from the perspective of word types and coined words, and future issues were also touched upon. Finally, we argue for the active use of high-quality open data.

View full abstract

This paper explains the essential knowledge required to understand the principles of multivariate analysis (including quantification theory). It introduces three representative methods: multiple regression analysis, cluster analysis, and correspondence analysis. For each method, the paper provides specific examples demonstrating their application in language research. Furthermore, it discusses points of caution when using these methods and mentions statistical software and related books to facilitate those attempting multivariate analysis.

View full abstract